Objective: To evaluate the interobserver agreement of ultrasonographic assessment of finger and toe joints in patients with rheumatoid arthritis (RA) by 2 investigators with different medical backgrounds.
Methods: Ultrasonography and clinical examination were performed on 150 small joints of 30 patients with active RA. A General Electric LOGIQ 500 ultrasound unit with a 7-13-MHz linear array transducer was used. In each patient, 5 preselected small joints (second and third metacarpophalangeal, second proximal interphalangeal, first and second metatarsophalangeal) were examined independently on the same day by 2 ultrasound investigators (an experienced musculoskeletal radiologist and a rheumatologist with limited ultrasound training). Joint effusion, synovial thickening, bone erosions, and power Doppler signal were evaluated in accordance with an introduced 4-grade semiquantitative scoring system, on which the investigators had reached consensus prior to the study.
Results: Exact agreement between the 2 observers was seen in 91% of the examinations with regard to bone erosions, in 86% with regard to synovitis, in 79% with regard to joint effusions, and in 87% with regard to power Doppler signal assessments. Corresponding intraclass correlation coefficient values were 0.78, 0.81, 0.61, and 0.72, respectively, while unweighted kappa values were 0.68, 0.63, 0.48, and 0.55, respectively. Ultrasonography showed signs of inflammation in 94 joints, while clinical assessment revealed tenderness and/or swelling in 64 joints.
Conclusion: An experienced radiologist and a rheumatologist with limited ultrasound training achieved high interobserver agreement rates for the identification of synovitis and bone erosions, using an introduced semiquantitative scoring system for ultrasonography of finger and toe joints in RA. Signs of inflammation were more frequently detected with ultrasound than with clinical examination. Ultrasonography may improve the assessment of RA patients by radiologists and rheumatologists.